writing to understand

writing doesn't record what you think. it discovers what you think. the act of putting words down changes the ideas themselves.

the discovery process

joan didion said it plainly: "i write entirely to find out what i'm thinking, what i'm looking at, what i see and what it means." paul graham expanded on this — he estimates that half the ideas in any essay are ones he thought of while writing it. not before. during.

this matches my experience exactly. i'll sit down to write about something i think i understand, and within a few paragraphs i'll hit a point where the sentence i'm writing doesn't quite work. the idea i had in my head was fuzzy, and the sentence demands precision. in the gap between the fuzzy idea and the precise sentence, new understanding forms.

this is not a pleasant process. it's closer to rubber duck debugging than to transcription. you're explaining your thinking to the page, and the page reveals that your thinking has holes.

why it works

paul graham puts it well: "if writing down your ideas always makes them more precise and more complete, then no one who hasn't written about a topic has fully formed ideas about it." that's a strong claim but i think it's right.

thinking feels fluid and complete in your head. you can hold a vague impression and feel like you understand something. writing forces serialization — you have to put one idea before another, make explicit connections, define terms. the constraints of language expose the gaps in thought.

it's the same mechanism as assumption auditing. you think you know how the system works until you have to explain it step by step. the explaining surfaces the broken assumptions.

my experience with it

i've written over a hundred reflections across different platforms — long-form in ideaflow, quick notes in apple notes, weekly reflections, blog drafts. the pattern is consistent: the writing that taught me the most was writing where i didn't know what i'd say when i started.

the reflections where i sat down with a clear thesis and just wrote it out — those were fine. useful for cementing what i already knew. but the reflections where i started with "i'm confused about X" or "something felt off about Y" — those were the ones that actually changed how i think.

the prerequisite

you have to be willing to write badly. if you're editing while writing, you're not discovering — you're performing. the discovery happens in the messy first draft where you follow an idea wherever it goes, even if it goes somewhere dumb. especially if it goes somewhere dumb. because "dumb" ideas, articulated, sometimes turn out to be not dumb at all — or they reveal something about what you actually think that a polished version would hide.

this connects to confidence. writing to understand requires the confidence to have bad ideas on the page. if you're afraid of writing something wrong, you'll stick to things you already know, and the discovery process doesn't happen.

the gap

the frustrating part: writing reveals insights, but insights don't automatically change behavior. see the-reflection-gap. understanding what you think is necessary but not sufficient for acting on it. the question is always: what do i do with what i just discovered?

writing vs. talking

talking also helps you think. but writing has an advantage: it's persistent. you can come back to it. you can notice contradictions between what you wrote last month and what you wrote today. you can watch your thinking evolve over time.

talking is better for speed and for the social dimension — someone can push back in real time, ask the question that cracks open a new line of thinking. but for deep, sustained, honest thinking, writing is better. you can't bullshit a page the way you can bullshit a conversation.